New Methods for Marginalization in Lazy Propagation
نویسنده
چکیده
Even though existing algorithms for belief update in Bayesian networks (BNs) have exponential time and space complexity, belief update in many real-world BNs is feasible. However, in some cases the efficiency of belief update may be insufficient. In such cases minor improvements in efficiency may be important or even necessary to make a task tractable. This paper introduces two improvements to the message computation in Lazy Propagation (LP). We introduce one-step lookahead methods for sorting the operations involved in a variable elimination using Arc-Reversal (AR) and extend LP with the anyspace property. The performance impacts of the methods are assessed empirically.
منابع مشابه
Solving CLQG Influence Diagrams Using Arc-Reversal Operations in a Strong Junction Tree
This paper presents an architecture for solving conditional linear-quadratic Gaussian (CLQG) influence diagrams (IDs) by Lazy Propagation (LP). A strong junction tree (SJT) is used to guide the elimination order whereas the marginalization operation is based on arc-reversal (AR). The use of AR for marginalization simplifies the implementation and gives the architecture a number of advantages. T...
متن کاملInference in Multiply Sectioned Bayesian
As Bayesian networks are applied to larger and more complex problem domains, search for exible modeling and more eecient inference methods is an ongoing eeort. Multiply sectioned Bayesian networks (MSBNs) extend the HUGIN inference for Bayesian networks into a coherent framework for exible modeling and distributed inference. Lazy propagation extends the Shafer-Shenoy and HUGIN inference methods...
متن کاملSemantic Learning for Lazy Clause Generation
Lazy clause generation (LCG) combines finite domain (FD) propagation with features from Boolean satisfiability (SAT) solving, notably clause learning. In a LCG solver, finite domain propagators are considered as clause generators that create a SAT description of their behavior for a SAT solver, using some Boolean encoding of the variable domains. When a conflict is triggered by unit propagation...
متن کاملOn Directed and Undirected Propagation Algorithms for Bayesian Networks
Message-passing inference algorithms for Bayes nets can be broadly divided into two classes: i) clustering algorithms, like Lazy Propagation, Jensen’s or Shafer-Shenoy’s schemes, that work on secondary undirected trees; and ii) conditioning methods, like Pearl’s, that use directly Bayes nets. It is commonly thought that algorithms of the former class always outperform those of the latter becaus...
متن کاملInference in Multiply Sectioned Bayesian Networks with Extended Shafer-Shenoy and Lazy Propagation
As Bayesian networks are applied to larger and more complex problem domains, search for flexible modeling and more efficient in ference methods is an ongoing effort. Mul tiply sectioned Bayesian networks (MSBNs) extend the HUGIN inference for Bayesian networks into a coherent framework for flexible modeling and distributed inference. Lazy propagation extends the Shafer-Shenoy and HUG IN infer...
متن کامل